Steep bill for AI buildout punishes Meta, other tech favorites

Meta’s announcement made the cost of the AI buildout more visible. The company plans to issue shares to fund infrastructure and other priorities, while also increasing its spending forecast. Investors interpreted that as evidence that even major technology companies may need outside capital to keep expanding. They also questioned how quickly the spending will produce returns.

Meta shares fell 8%, SpaceX declined 9%, and Nvidia dropped 4%. Meta reported negative free cash flow of $2.12 billion in the third quarter, compared with positive free cash flow of $10.63 billion a year earlier. That cash burn made new financing appear more important. New shares can also dilute existing holders, adding pressure to prices.

The selling spread across AI-related companies and semiconductors. Lam Research fell 12%, while the semiconductor index dropped 6.5%. Investors moved into Treasury bonds as a safety trade, reflecting concern that a pullback in AI spending could weaken US growth.

Meta’s Class C shares are a new category of stock with no voting power. Meta said it would use them to fund AI spending and other priorities. Their key feature is that investors can buy an economic interest in the company without receiving votes on corporate matters. The article does not describe their other specific terms.

When Meta issues these shares, investors pay money to the company in exchange for stock. Meta can then use that capital for data centers, other AI infrastructure, and related commitments. This approach is especially relevant because Meta reported negative free cash flow in the third quarter. The company therefore cannot rely entirely on internally generated cash for its planned expansion.

The trade-off is a larger number of shares outstanding. Even without voting rights, the new shares represent claims on Meta’s future value. Investors worried that repeated stock sales could dilute existing holders and signal that AI spending is exceeding current cash generation.

Meta’s planned spending is enormous. It raised the low end of its annual spending forecast from $130 billion to $135 billion, while leaving the high end unchanged at $145 billion. The company said this spending will support AI and other priorities. The scale matters because it increases the amount of financing needed before investors can judge the project’s eventual returns.

Meta’s cash position also changed sharply. In the third quarter, it reported negative free cash flow of $2.12 billion. A year earlier, it had reported positive free cash flow of $10.63 billion. Free cash flow shows whether the company generated or used cash after its spending, so the swing pointed to substantial cash use.

That combination worried investors. Meta may need to involve shareholders in future stock sales to fund additional building. The article describes the third-quarter result as Meta’s first cash burn in more than a decade, making the financing challenge especially notable.

Selling new shares can pressure existing shares because the company’s ownership is divided among more units. Each current share then represents a smaller percentage of the business, unless the new capital produces enough additional value to offset that change. Investors often react before those future benefits are proven.

Meta’s proposed Class C shares illustrate the mechanism. Meta would receive cash from buyers and use it for AI infrastructure and other priorities. But the company would also have more shares outstanding. The article says Wall Street expects companies such as Meta to sell billions of dollars in stock, especially as higher interest rates make borrowing less attractive.

The immediate effect can be weaker demand for existing shares. Investors may worry about dilution, future fundraising, and whether the spending will generate sufficient revenue. Meta’s negative free cash flow of $2.12 billion added to those concerns, because it suggested current operations were not fully funding the expansion.

Investors are questioning whether today’s AI infrastructure spending will generate enough profit, and when. Building out the technology requires large commitments before companies can clearly measure revenue from them. The article says the project is long term, its payoffs remain uncertain, and returns will likely be unevenly distributed.

Meta’s results made that concern concrete. It raised its annual spending forecast to as much as $145 billion and reported negative free cash flow of $2.12 billion in the third quarter. Investors saw more spending, cash burn, and insufficient revenue from that spending as signs that the buildout could become a money pit. Meta’s credit default swaps also reached a record, reflecting financing uncertainty.

The concern extends beyond Meta. Smaller AI companies may face much higher financing costs, while established hyperscalers can more easily raise money. If investors cannot identify who will capture the profits, they may sell AI-related shares and move into safer assets such as Treasuries.

Higher interest rates raise the cost of borrowing for companies that need debt to build infrastructure. AI firms may need financing well before their projects produce enough revenue to repay it. Rising credit risk adds another burden: lenders and bond investors may demand higher interest rates because they see a greater chance of financial stress.

The article points to smaller AI companies as especially vulnerable. CoreWeave’s unsecured bond yields soared to 15% in the previous week’s AI selloff. Nebius announced a new $3.5 billion credit line from Citigroup, while Bank of America analyst Tal Liani said CoreWeave may need to raise $20 billion to fund customer commitments.

Large companies such as Meta, Amazon, and Google are viewed as more capable of obtaining financing. But if funding becomes prohibitive for smaller firms, investors may become nervous without an outright collapse. Higher funding costs can slow construction, pressure share prices, and expose weaknesses in the AI buildout.

Data centers are facilities that house the computing equipment and systems needed to run digital services. AI infrastructure is the wider network of facilities, equipment, financing, and other support needed for AI activity. The article focuses on hyperscalers spending billions to build data centers and infrastructure that support the AI boom.

The upfront bill arrives before returns are clear. Companies must commit money to construction and capacity, then fund operations and customer commitments. Meta’s annual spending forecast is $135 billion to $145 billion, while its free cash flow turned negative in the third quarter. That means expansion can consume cash before it produces enough revenue to cover the investment.

This timing creates financing risk. Firms may borrow heavily or issue shares to continue building. Higher rates make debt more expensive, and stock issuance can dilute existing holders. Smaller companies face greater pressure because investors question whether they can secure enough funding while AI demand and profits remain uncertain.